In order to develop a preclinical model of liver-directed gene therapy in nonhuman primates, we compared the efficacy of 2 different vectors, 2 modes of delivery, 2 reporter genes and of immunosuppression in 8 rhesus monkeys. We inoculated 2 monkeys with an adenovirus vector containing the lacZ gene as a reporter and compared these with 2 other monkeys inoculated with a lacZ-containing plasmid in liposomes. Two from each group were inoculated via a mesenteric vein and 2 via the bile duct. The adenovirus resulted in much more gene expression than the liposome vector, and mesenteric vein inoculation was far superior to bile duct administration. The monkeys developed an immune response to the adenoviral vector, resulting in hepatitis and loss of transgene expression within 2 weeks. We immunosuppressed 2 monkeys with steroids and cytoxan and inoculated them with the adenoviral vector, 1 via the saphenous vein and 1 via a mesenteric vein. Both routes were equally effective, and there was prolonged transgene expression and no hepatitis. We next immunosuppressed 2 monkeys for only 2 weeks and inoculated them via the saphenous vein with adenovirus containing either lacZ or luciferase as a reporter gene. Hepatitis and loss of transgene expression occurred as soon as the immunosuppressive drugs were discontinued. Both reporter genes were easily detected. We plan to continue these studies with more advanced vectors and more refined methods of immunosuppression in order to develop a good model for preclinical testing of products for gene therapy.